ROCm / AITemplate

AITemplate is a Python framework which renders neural network into high performance CUDA/HIP C++ code. Specialized for FP16 TensorCore (NVIDIA GPU) and MatrixCore (AMD GPU) inference.
Apache License 2.0
11 stars 7 forks source link

make merge_upstream branch support gfx1100 #73

Closed Boom-Hacker closed 12 months ago

Boom-Hacker commented 12 months ago

I tested,sd compile pass,but it noticed this

"pipeline_utils.py:818: RuntimeWarning: invalid value encountered in cast images = (images * 255).round().astype("uint8")"

and get a pure black photo